Updated: 9 August 2010

Use NORMINV to calculate the inverse of the normal cumulative distribution function (cdf). The formula for cumulative distribution function is:

SELECT [wctStatistics].[wct].[NORMINV] (

<@Probability, float,>

,<@Mean, float,>

,<@Standard_dev, float,>)

is a probability corresponding to the normal distribution. *@Probability *is an expression of type **float** or of a type that can be implicitly converted to **float**.

is the arithmetic mean of the distribution. @*Mean *is an expression of type **float** or of a type that can be implicitly converted to **float**.

is the standard deviation of the distribution. @*Standard_dev *is an expression of type **float** or of a type that can be implicitly converted to **float**.

float

· If @Probability < 0 or @Probability > 1, NORMINV returns an error

· If *@Standard_dev* = 0, NORMINV returns an error

· If @*M**ean* = 0 and *@Standard_dev *= 1, NORMINV = NORMSINV(*@Probability*)

· NORMINV uses root-finding algorithms and iteration to solve for NORMDIST(NORMINV(*@Probability*, *@Mean, @Standard_dev*), *@Mean, @Standard_dev*,'True') - *@Probability *= 0to eight decimal places

Select wct.NORMINV(0.00173238078965338,10,5)

This produces the following result

----------------------

-4.61589999992911

(1 row(s) affected)

Select Round(wct.NORMDIST(wct.NORMINV(0.00173238078965338,10,5),10,5,'True')-0.00173238078965338, 9)

This produces the following result

----------------------

0

(1 row(s) affected)